Transforming a Customer Data Ecosystem on AWS


Transforming a Customer Data Ecosystem on AWS

Major Theme Parks Company

Transforming a Customer Data Ecosystem on AWS

Major Theme Parks Company

Transforming a Customer Data Ecosystem on AWS

Major Theme Parks Company

Transforming a Customer Data Ecosystem on AWS

Major Theme Parks Company

Transforming a Customer Data Ecosystem on AWS

Transforming a Customer Data Ecosystem on AWS

Overview & key takeaways

A major theme parks entertainment company sought to build a 360-degree view of its guests to drive more personalized marketing campaigns. Together with Merkle they integrated guest data from disparate marketing systems and transformed its guest view from transaction-centric to person-centric.

  • 150M

    records processed from 15 different raw feeds

  • 10M

    records enriched with additional demographic attributes

  • Ability to analyze 20 million identified guests on an annual basis

  • Minimum viable product deployed into production within 3 months

The challenge

A major theme parks company needed to integrate its guest data from disparate marketing systems to create a 360-degree view of customers, building customer profiles that compile all data together as it’s collected. Additionally, the park’s data quality was lacking due to its systems only providing an acute view of guests, with fractioned pieces of information like email address, and first and last name. However, there wasn’t enough information to provide a full view of each guest. Previous attempts to create such holistic profiles were met with limited success due to lack of a scalable, highly performant, and economic computing platform.


The approach

Merkle and the client mutually concluded that Amazon Web Services (AWS) would be the ideal platform to build this proposed solution. AWS provides fully managed storage, computing, and analytics services on its platform, thereby resulting in faster time to value and ease of development. 

Merkle helped the theme park company integrate all of the guest data, transactions, and interactions onto an AWS-based cloud-native solution while enabling its data science team to segment guests more granularly to create personalized offers. During this process, Merkle developed a marketing data platform on AWS that allowed the theme park company to ingest guest data from multiple source systems, resolving guest identity across systems to link them together. The cloud- based environment served as the focal point for data consumption by various teams across the client’s organization in understanding guest activity and experience in the park, to segment guests for marketing purposes, import marketing lists into a campaign tool to build personalized offers, and derive analytical insights such as guest propensity to visit parks, perform certain activities, etc. 

To enhance existing data quality, Merkle generated global unique identifiers for guests, standardized and cleansed all personally identifiable information (PII) through various Merkle proprietary tools under its Merkury suite of solutions. Having a global identifier associated with each guest allowed the park to connect guest activity across disparate systems and generate insights which were previously not possible. Merkle also leveraged its data asset called DataSource for reverse email and phone appends to create a more complete PII profile of individuals. DataSource also helped enrich the PII profile with demographic elements which were previously not available to the client, such as age, household income, household size, etc. These additional elements provided important information to granularly segment audiences for personalized campaigns.


The outcome

Merkle used the following AWS services to build the solution:

  • AWS Redshift: Cloud-based database to build and host guest 360 view 
  • AWS Elastic Computing Cloud: Compute machines to build Python based data pipelines using Merkle proprietary Data Loading Framework (DLF) 
  • AWS Simple Storage System: Landing area for raw data acquired from disparate source systems 
  • Elastic Container Service: Apache Airflow containers for orchestration of data pipelines 
  • Relational Database Service: Postgres database to store metadata from Apache Airflow and Microsoft SQL Server for data pipeline metadata